Lieutenant General Yoshitsugu Saito had finished his final orders, and what he wrote was not a tactical assessment. It was a statement of purpose. Every man who could hold a rifle should join the final attack. There was to be no surrender.
And those who could not fight should die by their own hand rather than fall into American hands. He had been on Saipan for months, had watched the Japanese airpower over the Marianas erode and vanish, had known for some time that the relief force promised by planning would never come. The naval battle fought in June, the one the Americans would call the Great Marianas Turkey Shoot, had destroyed the carrier air groups that were supposed to provide cover for a counteroffensive. He commanded a garrison that could not be reinforced, could not be supplied, could not be relieved.
Now he was telling his men to die in a particular way, for a strategic purpose they could not fully understand. Saito followed his own orders. On July 6th, 1944, before the final banzai charge he had commanded was executed, he died by his own hand. The island he left behind lay 1,400 miles south of Tokyo.
Saipan was not a distant coral atoll whose strategic value rested on air range and supply lines. It was not a jungle island whose garrison could be isolated and left to wither. It was within B-29 range of the Japanese home islands. When American forces took Saipan, American bombers could reach Tokyo.
The implications for the Japanese High Command were not ambiguous. Losing Saipan meant losing the air defense buffer that stood between the American bomber force and the cities of Japan. The island also held a civilian population of approximately 25,000 Japanese and Korean civilians, brought there in peacetime as agricultural workers, teachers, merchants, children, and by June of 1944 thoroughly embedded in the fabric of the garrison’s daily life and completely subject to the authority of the military command. The American assault began on June 15th, 1944, with a naval bombardment the defenders had been anticipating for weeks.
The 2nd and 4th Marine Divisions hit the beaches on the southwest coast simultaneously, supported by the most concentrated naval gunfire yet employed in a Pacific amphibious operation. The landing was not unopposed. Japanese artillery covered the beaches, machine gun nests were built into the coral ridges above the landing zones, and mortar positions were registered on the most likely approach routes. The first day cost the Marines nearly 2,000 casualties, killed and wounded, the most expensive single day of amphibious fighting in the Pacific war to that point.
Japanese counterattacks that night, including a tank assault that hit the Marine perimeter in darkness, added to the cost on both sides and demonstrated that the garrison intended to contest every phase of the battle. Saito had ordered precisely the kind of defense that General Tadamichi Kuribayashi would later refine on Iwo Jima. No beach defense. No wasteful banzai charges in the opening phase.
A fighting withdrawal designed to make every yard of advance cost the Americans something in blood. The terrain cooperated. Saipan’s limestone ridges, caves, dense vegetation, and the volcanic spine of the island running north to south gave the defenders a series of natural positions that each had to be taken individually at close range against men who were not going to withdraw. The north-south ridge system known as Mount Tapochau and its surrounding heights provided the layered defensive ground that rewarded the defender and punished the attacker in exactly the way Saito needed to make the battle last as long as possible and cost as much as possible.
By the end of the first week, the American forces had secured a substantial beachhead and were pushing inland. The difference between Saipan and Iwo Jima was not the doctrine. It was the terrain, the timeline, and those 25,000 civilians, who were about to become part of the battle in a way that had no precedent in the Pacific war and no solution in any military doctrine either side had developed. As the advance pushed north through June into July, Saito communicated with Tokyo by radio.
Tokyo told him to hold, to fight to the last man, to make Saipan cost the Americans enough that an invasion of the home islands would be politically impossible. He issued his final orders to soldiers who had been fighting for three weeks without meaningful resupply, who were wounded in numbers that would have rendered most military forces non-functional, and who faced an American force with numerical superiority and complete air and naval fire support. The orders were consistent with the Senjinkun, the Japanese military code that defined surrender as a disgrace worse than death, and with everything Japanese military culture had built into its men through training and doctrine. On the morning of July 7th, 1944, approximately 4,000 Japanese soldiers moved south against the American lines in the final organized military action on Saipan.
It was the largest banzai charge of the entire Pacific War. The men who participated were not, in most cases, soldiers in any conventional operational sense by that point in the battle. They were the survivors of three weeks of grinding defensive combat, wounded, starving, largely out of ammunition. Many were armed only with bayonets fixed to rifles that could no longer fire, or with sharpened bamboo poles.
They had been told to die fighting rather than surrender, and the charge was the form that dying fighting took when there was no longer any tactical situation to manage or any defensive position to hold. The American units the charge hit, primarily Army battalions that had replaced Marine units in the northern sector, were initially overwhelmed by the sheer volume of men moving through and past their positions. The charge penetrated deeply enough to reach rear area units, including artillery crews, who fought with small arms and whatever was at hand against men who had broken through the forward lines. The fighting lasted for several hours and cost the Americans heavy casualties before reinforcing units sealed the penetration and the last organized resistance collapsed.
When it was over, approximately 4,000 Japanese soldiers were dead in a stretch of ground the Americans called Harakiri Gulch. The charge had accomplished nothing militarily that could change the outcome of the battle. That was not why it had been ordered, or why the men who carried it out had done so. It had been ordered as the specific form of death that Japanese military culture had defined as honorable, and the men who died in it had accepted that definition fully enough to walk into American fire with weapons that could not hurt anyone.
There is something in the account that gets less attention. The soldiers who died in the charge on July 7th were combatants. They had been trained, however inadequately in the end, for the possibility of dying in battle. The 25,000 Japanese and Korean civilians on Saipan had not been trained for anything.
They had come to the island in peacetime to build lives in a settlement the Japanese government had been developing for decades. In the final days of the battle, as organized resistance collapsed and American forces reached the northern cliffs, somewhere between 1,000 and 10,000 of those civilians, the exact number has never been established with certainty, jumped from the cliffs into the ocean or onto the rocks below rather than surrender. They did this because Japanese military and civilian authorities had told them for years that Americans were barbarians who would torture, rape, and murder any Japanese they captured. They did this because the military commanders on the island had distributed grenades to civilian groups and told them to use those grenades on themselves rather than be taken alive.
They did this because the ideological framework of wartime Japanese nationalism had constructed the relationship between the Japanese people and the enemy in terms that made death seem more rational than capture. American soldiers who watched from a distance and tried to call out in Japanese through interpreters that surrender was safe, that no harm would come to them, were ignored, or not understood, or not believed. The cliffs are called Banzai Cliff and Suicide Cliff in postwar accounts. They are among the most documented mass civilian deaths of the Pacific War, and they are entirely the product of a government telling its people a lie about the enemy that those people believed completely enough to die rather than discover it was wrong.
Against this backdrop, the approximately 1,000 Japanese soldiers and civilians who surrendered or were captured on Saipan represent a different story, one the Japanese military command had specifically tried to prevent and American intelligence was extremely interested in documenting. The prisoners came from a wider range of backgrounds than the garrison’s combat arms units. Some were support troops cut off from their units in the chaotic final days. Some were wounded men physically incapable of the resistance their training required.
Some were men who, in the isolation and confusion of the final days, had arrived at a private decision that the doctrine was wrong, that dying on an island the war had already abandoned was not the service to Japan the code described. None of them surrendered easily or without the specific kind of psychological crisis the doctrine was designed to prevent. The cultural architecture that produced the deaths on the cliffs operated on the prisoners too. What made them different from the men who died was not the absence of that architecture, but the specific circumstances of their capture.
They were processed through the Joint Intelligence Center Pacific Ocean Area, where trained interrogators who understood Japanese language and culture spent weeks in careful conversation with men who had just survived an experience that virtually every aspect of their cultural conditioning had told them would end in death or in something worse than death. What the interrogators found was a population of survivors whose experience of American captivity immediately began dismantling the picture of Americans that Japanese propaganda had built. The interrogators were not brutal. The conditions of captivity were not what the prisoners had been told to expect.
Food was provided. Medical treatment was given to the wounded. Men who expected to be killed or tortured found themselves alive and physically cared for in circumstances that directly contradicted everything the official narrative had told them about what happened to Japanese who fell into American hands. The prisoner accounts of the civilian deaths produced some of the most disturbing testimony of the Pacific War, and the most disturbing element was not the deaths themselves but the specific chain of causation the survivors described.
The civilians who jumped from the cliffs had been told two things with complete confidence by Japanese military and civilian authorities. First, that Americans routinely tortured and killed Japanese prisoners. Second, that the Japanese military would return to Saipan and liberate any Japanese who survived American occupation, but that anyone who had cooperated with American forces or been in American captivity would be treated as a traitor when that liberation came. Both claims were false.
The prisoners who described them knew the first claim was false because they were sitting in American captivity in conditions that contradicted it directly. The second was rendered irrelevant by the strategic situation that had produced Saipan’s fall in the first place. But the civilians on the cliffs had not yet had the opportunity to discover that the claims were false. They made their decisions on the basis of information their government provided in a cultural framework their government had constructed over years of systematic propaganda.
The decision to die rather than discover the truth was the product of that information and framework operating together in a moment of absolute crisis, with no time for deliberation and no available corrective. The survivors who provided the testimony were largely matter-of-fact about what they described, the way people are matter-of-fact about things they cannot change and have already processed to whatever extent processing is possible. The fall of Saipan produced a political earthquake in Tokyo that had no precedent in the Pacific war. Within a month of the island’s fall, Prime Minister Hideki Tojo had resigned and the Japanese government underwent a restructuring that, while it did not change the fundamental direction of the war, represented the first serious political consequence of the military situation in the Pacific reaching the level of the civilian government.
Tojo’s resignation was directly caused by the loss of Saipan and what that loss meant for the strategic position of the home islands. The men who replaced him understood, as Tojo had, that the war was going badly. They also understood that acknowledging publicly what they understood privately was politically and institutionally impossible. The Japanese public in July of 1944 was still being told that the war was proceeding in a way consistent with ultimate Japanese victory.
The civilians who died on Saipan’s cliffs had believed this. The soldiers who died in the banzai charge had believed variations of it. The political system that had produced these beliefs could not, without destroying itself, announce that the beliefs were wrong. So it did not announce anything of the kind.
It replaced Tojo with other men, restructured the military command, and continued the war on the assumption that something would change. That American losses in the coming battles for Peleliu, the Philippines, Iwo Jima, and Okinawa would break American political will. That home island defense would prove so costly that a negotiated peace would become available. The thousand men who surrendered on Saipan, sitting in American captivity and describing to interrogators the difference between what they had been told and what they found, represented the truth the Japanese government could not afford to tell and could not prevent those men from eventually telling.
The Battle of Saipan decided more than the strategic question of B-29 range to Japan. It decided the framework within which the final year of the Pacific war would be fought. The Japanese military’s response, maintaining the no-surrender doctrine, intensifying the psychological preparation of soldiers and civilians for death rather than capture, and escalating kamikaze tactics that became formalized doctrine in October of 1944, was the response of an institution that had concluded that changing the terms of the war was impossible and that the only available strategy was to make American victory so costly that the terms of peace would be better than unconditional surrender. The American military’s response to the civilian deaths on the cliffs and to the intelligence picture the prisoners produced was to begin planning for a home island invasion on the assumption that the Japanese population had been conditioned to resist in exactly the way the civilians on Saipan had resisted, and that an invasion of Japan would produce civilian casualties on a scale that made Saipan look like a preliminary.
The thousand who surrendered had shown American planners the mechanism by which Japanese civilian resistance was produced. Specific lies about American behavior delivered with institutional authority in a cultural framework that made death seem rational. Understanding the mechanism did not make the problem easier to solve. It simply made it more clearly visible.
The decisions made in response to that visibility, the atomic bombs, the Soviet declaration of war against Japan, the surrender that followed, were all downstream of what Saipan had shown both sides about the nature of the war they were fighting and about how it could end. Thirty thousand Japanese soldiers defended Saipan. Fewer than 1,000 chose to surrender. The rest died on the island, died in the charge, or died on the cliffs.
The thousand who surrendered sat in American captivity and described to men with notebooks the difference between what they had been told and what they found. That difference was the truth of Saipan. The government that produced the deaths on the cliffs understood it perfectly.
It was the people who died who did not get to find out.
